Research

Widely taught trading strategies, coded rule by rule, compared with a simple control and, where possible, rerun on data nobody had looked at. Every number comes from a results file and can be recomputed.

• Study no. 2 · November 8, 2026

  Five SMC strategies. None beats buying a pullback.

  Five SMC/ICT setups coded rule by rule and tested on two years of gold M1, V75 and five synthetic indices: none does better than simply buying a pullback in the trend (p from 0.41 to 0.998).

  Read the study →
• Study no. 1 · November 8, 2026

  Every internal check said yes. Fresh data said no.

  A candlestick setup passes every check on the data where it was found (+11.30 points per trade), then loses on 374 never-seen days (−4.89 points per trade).

  Read the study →